Publication number | US7982799 B2 |
Publication type | Grant |
Application number | US 11/618,337 |
Publication date | Jul 19, 2011 |
Filing date | Dec 29, 2006 |
Priority date | Dec 30, 2005 |
Fee status | Paid |
Also published as | DE102005063072A1, EP1804492A2, EP1804492A3, US20070153124 |
Publication number | 11618337, 618337, US 7982799 B2, US 7982799B2, US-B2-7982799, US7982799 B2, US7982799B2 |
Inventors | Ulrich Niedermeier |
Original Assignee | Trident Microsystems (Far East) Ltd. |
Export Citation | BiBTeX, EndNote, RefMan |
Patent Citations (27), Non-Patent Citations (5), Referenced by (2), Classifications (11), Legal Events (6) | |
External Links: USPTO, USPTO Assignment, Espacenet | |
This patent application claims priority from German patent application 10 2005 063 072.3 filed Dec. 30, 2005, which is hereby incorporated by reference.
The invention relates to the field of video signal processing, and in particular to interpolation of an image information value for a pixel of an interline situated between two original image lines in an image.
For the transmission of video images or television images, it is known how to transmit interlaced fields instead of frames in which an image information value is present for each pixel of the image. Each of these frames transmitted contains a number of scanning lines, each of which contain a number of pixels. Image information values, i.e., luminance values or chrominance values, are only present for the pixels of every other line, and frames are transmitted alternatingly in which image information values are present for pixels of even numbered lines and for pixels of odd numbered lines.
In order to turn such a transmitted field into a frame for representation on a display device, such as a monitor screen, it is necessary to interpolate image information values for the interlines. In so-called intra-field algorithms, the interpolation of an image information value for a pixel of an interline of a field only uses image information values of pixels from the same field.
In the most elementary case, the available image lines of a field are doubled to produce a frame. The image information values of the pixels of an interline interpolated in this way then correspond to the image information values of the image lines lying above or below the interline. Edges running in the diagonal direction through an image being depicted will, however, appear as “steps” in the frame with this kind of interpolation. One method for avoiding such steplike artifacts in a frame generated from a field is described, for example, in U.S. Pat. No. 5,625,421.
In the Edge Based Line Average (ELA) algorithm, for each pixel to be interpolated one determines the direction of a possible edge containing this pixel. The algorithm is described, for example, in T. Doyle, “Interlaced to Sequential Conversion for EDTV Applications” Proc. of 2nd Int. Workshop on Signal Processing of HDTV, L'Aquila, Italy, 1988, or in T. Doyle, and M Looymans, “Progressive Scan Conversion Using Edge Information” Proc. of 3rd Int. Workshop on HDTV, Torino, Italy, 1989, or in Lee et al.: “A New Algorithm for Interlaced to Progressive Scan Conversion Based on Directional Correlations and its IC design,” IEEE Transactions on Consumer Electronics, Volume 40, Number 2, May 1994, page 119.
In the ELA algorithm, one determines the difference, for various image directions, between the image information values of two pixels lying adjacent to the pixel being interpolated in the particular direction. The direction for which the magnitude of this difference is a minimum is used as the direction of the edge contour. The image information value of the pixel being interpolated is then interpolated by making use of pixels situated adjacent in this direction to the pixel being interpolated. Ambiguities can present a problem in this method, as will be explained hereinbelow with reference to
In
Other methods of interpolation of an image information value are described in U.S. Pat. No. 6,965,705, EP 0 550 231, published U.S. Patent Applications 2005/0157951 and US 2005/0134602, PCT application WO 99/19834, or in FAN, YU-CHENG; et al.: “Intelligent Intra-Field Interpolation for Motion Compensated Deinterlacing”, 3rd International Conference on Information Technology ITRE 2005. 27-30 Jun. 2005, pp. 200-203.
There is a need for an interline interpolation that ensures an improved interpolation technique of edges running through the image.
Interpolation of an image information value for a pixel of an interline calls for selecting, from a number of image directions, to each of which a direction quality value is assigned, a direction of interpolation by selecting these direction quality values, and interpolating the image information value being interpolated in dependence on image information values assigned to pixels lying adjacent to the pixel being interpolated in the direction of interpolation. The following steps of the method are specified for determining a direction quality value for an image direction:
determine a single direction quality value for each pixel of a pixel group with at least two pixels, the single direction quality value being dependent on image information values assigned to image regions lying adjacent to the particular pixel in the image direction,
create the direction quality value as a function of the single direction quality values (q) of the pixel group.
The direction quality value may correspond to the mean of the single quality values.
A single direction quality value assigned to an image direction represents a measure of the probability that an edge travels in the particular image direction in the image. Now, in order to determine, for a pixel being interpolated, the probability of the presence of an edge contour in a given image direction, several such single direction quality values, representing a probability of the presence of an edge contour, are added up and the thus obtained (composite) direction quality values are used to select an image direction as the direction of interpolation. Investigations have shown that such an addition of the single direction values of several adjacent pixels can prevent or at least reduce ambiguities in the determination of an optimal image direction.
These and other objects, features and advantages of the present invention will become more apparent in light of the following detailed description of preferred embodiments thereof, as illustrated in the accompanying drawings.
Interpolation of an image information value of a pixel of an interline of a field shall now be explained.
The original image lines y−3, y−1, y+1, y+3 with the original pixels and the interlines y−2, y, y+2 being interpolated, with the pixels being interpolated, are part of an image in which the original pixels and the pixels being interpolated are arranged in a matrix. A first direction of this pixel matrix, in which the pixels of a line are arranged adjacent to each other, is termed hereinafter the horizontal image direction or x direction, while a second direction in which the individual lines are arranged adjacent to each other is termed hereinafter the vertical direction or y direction.
The original pixels (i.e., the pixels of the original image lines) are each assigned image information values, while the pixels being interpolated (i.e., the pixels of the interlines) are to be assigned such image information values by interpolation of the image information values of original pixels. These image information values may be luminance values or chrominance values.
Image directions can be defined in the image depicted partially in
where dx denotes the horizontal distance between two pixels arranged adjacent to each other in the particular image direction (i.e., they lie on a straight line running in the image direction); dy denotes the vertical distance between two pixels lying on this straight line running in the image direction. An image direction in the following discussion for dy=1 is defined by using the horizontal distance dx of two pixels lying on the straight line running in the image direction. In this case we have:
dx=arctan(φ) (2)
where dx thus corresponds to the horizontal distance of two pixels lying in image lines arranged directly next to each other in the vertical direction, and lying on the straight line establishing the image direction. The vertical is defined as dx=0. Referring to
The definition of an image direction using the horizontal distance dx is, of course, not limited to integer values of dx. Thus, dx=0.5 denotes, for example, an image direction defined by two pixels whose horizontal distance is dx=1 and whose vertical is dy=2.
For interpolation of an image information value S(x,y) for the pixel 15 being interpolated, having coordinates (x,y), one will select a group of pixels, which may include the pixel being interpolated. Referring to
Direction quality values for different image directions for each of these pixels 11, 13, 15, 17, 19 are determined and summed to obtain composite direction quality values for the individual image directions for the pixel 15 being interpolated.
As can be seen from the plot of the single direction quality values in
It has been found that such ambiguities can be for the most part avoided by adding the single direction quality values of the pixels of a group being interpolated, to ascertain composite direction quality values for different image directions. In the example of
By comparing the direction quality values determined for the individual image directions and assigned to the pixel being interpolated one selects an image direction as the interpolation direction and an image information value is interpolated for the pixel 15 being interpolated by making use of original pixels lying adjacent to the pixel being interpolated in the direction of interpolation. The selection of an image direction as the direction of interpolation involves, for example, determining a minimum for the composite direction quality values and selecting the image direction for which the minimum exists as the direction of interpolation.
The interpolation of the image information value is done, for example, by forming the mean of the pixels of the original image lines y−1, y+1, which lie adjacent to the pixel 15 being interpolated in the direction of interpolation. In the following, we shall denote by dx0 the direction of interpolation found by means of the composite direction quality values, and we then have for the interpolated image information value of the pixel (x,y) being interpolated:
Whereas
We shall use below the following nomenclature for the single direction quality values: q(u,v)[dx], where q denotes here a single direction quality value, u,v denote the coordinates of the pixel to which this single direction quality value is assigned, and dx denotes the particular image direction for which the single direction quality value was determined. The nomenclature qd(u,v)[dx] will be used below for the composite direction quality values. For the composite direction quality value assigned to the pixel 15 being interpolated with the coordinates (x,y) we have:
where i and k denote here the coordinates of the pixels of the group relative to the coordinates of the pixel 15 being interpolated. For the example explained with reference to
Preferably, the group of pixels whose single direction quality values are added to determine a composite direction quality value is dependent on the particular examined image direction, as will be explained below with reference to
Referring to
A second group of pixels whose single direction quality values are used to determine a composite direction quality value for the image directions dx=−3, −2 and dx=2, 3 are represented in
In the above-explained method, each group of pixels includes a subgroup with a number—in the example, six—of the pixels arranged at a distance from the pixel 15 being interpolated in the horizontal direction. The horizontal distance of the pixels of this subgroup from the pixel 15 being interpolated is greater as the absolute magnitude of the value dx defining the image direction is larger. This is tantamount to saying that the horizontal distance of the pixels of this subgroup from the pixel being interpolated is greater as the angle φ dictating the image direction deviates more from the vertical, or the closer this angle is to the horizontal.
To reduce the computational expense, one preferably determines composite direction quality values not for every pixel of an interline being interpolated, but only those for every n-th pixel of an interline being interpolated. One then determines a direction of interpolation for each of these n-th pixels by the composite direction quality values and this direction of interpolation is used as well for the adjacent pixels in the horizontal direction. Referring to
As already explained, the single direction quality values found for a pixel represent a measure of a probability that an edge runs in the particular direction in the image, for which the single direction quality value has been found. Determining the single direction quality values for a pixel will now be explained by determining a single direction quality value for the pixel 15 with the coordinates (x,y). The finding of the single direction quality value q(x,y)[dx] explained for this pixel 15 applies equally to the other pixels of the group of pixels whose single direction quality values are used to find the direction quality value.
The single direction quality value is dependent on a direction correlation value dm(x,y)[dx]. This direction correlation value forms a measure of the correlation of the image information values of at least two pixels, situated in a given image direction dx adjacent to the pixel with the coordinates (x,y) for which the single direction quality value is being determined. In the most elementary case, the direction correlation value corresponds to the difference of the image information values of the pixels situated in the investigated image direction dx adjacent to the particular pixel, and so we have:
dm(x,y)[dx]=|S(x−dx,y−1)−S(x+dx,y+1)| (5).
|•| denotes an absolute value. The direction correlation value in this case corresponds to the difference of the image information values assigned to the pixels.
To determine the direction correlation value, making reference to
The variable δ here represents a measure of the number of additional image information values needing to be taken into account.
Generally, the direction correlation value dm(x,y)[dx] is a function of the absolute value of the difference in image information values of at least two pixels situated adjacent to each other in the examined image direction dx.
In general, the single direction quality value q(x,y)[dx] for the pixel 15 is dependent on the particular direction correlation value found for the pixel 15, and therefore:
q(x,y)[dx]=f(dm(x,y)[dx]) (6)
where f(•) here denotes a function mapping the direction correlation value onto the single direction quality value.
The single direction quality value q for a given direction may correspond, in particular, to the direction correlation value for the respective direction.
A direction correlation value found in the above explained manner for a particular direction dx is smaller as the image information values or the original pixels situated adjacent to each other in the examined image direction dx differ less from each other. Hence, the probability of the presence of an edge contour in this direction is greater the less the image information values of these original pixels differ from each other (i.e., the smaller is the direction correlation value).
Preferably, one will use the vertical direction to determine the direction correlation value, i.e., the direction with dx=0. This can be achieved by adding a constant or variable value to the direction correlation values dm of all the image directions outside of the vertical direction when determining the single direction quality value.
In one embodiment, a vertical correlation value is added to the direction correlation value of each direction. This vertical correlation value represents a measure of the deviation between an interpolation outcome obtained by interpolation in the vertical direction and the interpolation outcome obtained by interpolation in another direction deviating from the vertical direction. For these vertical correlation coefficients we have:
mv(x,y)[dx]=|(S(x−dx,y−1)+S(x+dx,y+1))−(S(x,y−1)+S(x,y+1))|=|ip(x,y[dx])−ip(x,y[dx=0])| (7).
The first term ip(x,y)[dx] here represents the mean of the original pixels situated in the image direction dx≠0 adjacent to the pixel with the coordinates (x,y), while the second term ip(x,y)[dx=0] represents the mean of the image information values of the pixels which are situated in the vertical direction adjacent to the pixel with the coordinates (x,y). For the single direction quality value q we have, using the vertical correlation value:
q(x,y)[dx]=k1·dm(x,y)[dx]+k2·mv(x,y)[dx] (8)
where k1, k2 denote here weighting factors which determine what weight to allot to the vertical correlation value when determining the single direction quality value. Since mv[dx=0]=0, the vertical direction is implicitly preferred when determining the single direction quality value. A direction other than the vertical direction will only furnish a smaller single direction quality value than the vertical direction if its direction correlation value in consideration of the weighting factors k1, k2 is smaller than the vertical correlation value mv.
The weighting factor k2 by which the vertical correlation value mv is multiplied is preferably variable, to prevent giving preference to the vertical direction in those cases where definitely or with high probability there is no edge contour in the vertical direction.
The probability of the presence of an edge in the vertical direction can be determined, for example, by investigating the image information values of the original image lines situated above and below the image line being interpolated. This will be explained with reference to
The above explained vertical weighting factor k2 is found by a comparison of the region values m1, m2. For example, for the vertical weighting factor k2 we have:
k2=k3·vg (9)
where k3 here represents a constant factor for the weighting of the vertical correlation coefficient mv, while vg is a value dependent on the comparison of the region values m1, m2, which can take on three different values, for example, namely 0, 1 or 2. For this variable vertical weighting factor we have, for example, vg=2 when none of the following conditions I to IV is fulfilled:
vg=0 for m1>4 & m2<−4 (I)
vg=0 for m1<−4 & m2>4 (II)
vg=0 for m1=6, m1=−6, m2=6, m2=−6 (III)
vg=1 for |m1|>2, |m2|>2, (IV).
when (I), (II), (III) not fulfilled
A comma (,) in conditions (I) to (IV) stands for a logical OR operation.
An antiparallel increase or decrease of the image information values in the original image lines y−1, y+1, such as when conditions (I) or (II) are obtained, constitutes strong evidence that no edge contour is present in the vertical direction. In this case, the vertical weighting factor is vg=0, so as not to give preference to the vertical direction when ascertaining the single direction quality value (e.g., see EQ. (8)). Such an antiparallel increase or decrease of the image information values can, instead, indicate a thin, slanted line. Such a line is an image structure with two parallel running edges.
In another embodiment, the conditions for the vertical weighting factor vg=1 are defined more narrowly and vg=1 only if none of the conditions (I) and (II) and one of the following conditions is fulfilled:
m1≧2& m2≦−1 (Va)
m1≧1 & m2≦−2 (Vb)
m1≦−2 & m2≧1 (Vc)
m1≦−1 & m2≧2 (Vc).
The information that no edge contour is present in the vertical direction can be used, arguing conversely, to give preference when determining the single direction quality value to those image directions which depart especially greatly from the vertical direction. This can be accomplished when determining the single direction quality value by adding to the term dependent on the direction correlation value dm a further term which is direction-dependent, and which is smaller as the particular direction deviates more from the vertical. Thus, for the single direction quality value q we have:
q(x,y)[dx]=k1·dm(x,y)[dx]+k3·vg·mv(x,y)[dx]+k4·fg·pf[dx] (10).
The term k4·fg·pf[dx] here represents a direction-dependent value. k4 is a constant weighting factor, indicating how strongly this direction-dependent value should be weighted in relation to the direction correlation value dm. The additional weighting factor fg is dependent on the vertical weighting factor vg, with:
fg=1 for vg=0 (11a)
fg=0 for vg≠0 (11b).
Thus, fg is only 1 to give preference to directions deviating from the vertical direction, when the variable vertical weighting factor vg is equal to zero, and thus when the presence of an edge contour in the vertical direction can be ruled out with high probability. The other factor pf[dx] is dependent on the particular image direction and will be termed hereinbelow the direction-dependent weighting factor. This direction-dependent weighting factor pf[dx] decreases the more the particular image direction deviates from the vertical direction.
pf[dx]=a−|dx| (12).
The larger the value dx determining the image direction here, the smaller the weighting factor pf[dx] and the more heavily preferred are “shallow image directions” in ascertaining the single direction quality value according to equation (10). The value a in equation (12) is a constant given value, which for example is equal to the value dx for the “shallowest” image direction investigated when determining the single direction quality value, i.e., the image direction whose direction angle φ is the largest.
It is possible to use a constant additive value pv[dx] when determining the single direction quality values for all directions except the vertical direction. We then have, for the determination of the single direction quality value:
q(x,y)[dx]=k1·dm(x,y)[dx]+k3·vg·mv(x,y)[dx]+k4·fg·pf(x,y)[dx]+pv[dx] (13).
For pv[dx] we have:
pv[dx]=0 for dx=0 pv[dx]=b for dx≠0 (14).
This technique ensures that preference is given to the vertical direction in cases where no significant minimum is found for any image direction, so that the vertical direction is chosen as the direction of interpolation.
Preferably, when determining pv[dx], we also allow for the vertical weighting factor vg, so that we have:
pv[dx]=0 for dx=0 and vg=2
pv[dx]=b for dx≠0 or vg=0,1 (15).
The advantage of this technique is that there is a preference for the zero direction only for the case vg=2, i.e., when no lengthier monotonic functions are present above or below, and thus there is slight probability for the presence of a thin, shallow line.
The arrangement has a first computation unit 103, which determines for the respective image direction dx the direction correlation value dm[dx]. A second computation unit 104 determines the above explained vertical correlation value mv[dx], and a third computation unit 105 computes the above explained variable vertical weighting factor vg. A fourth computation unit 106 provides the direction-dependent weighting factor pv[dx]. The individual computation units are provided with the information dx about the particular image direction being investigated.
A first multiplier 107, connected to the computation unit 103, multiplies the direction correlation coefficient dm[dx] by the constant weighting factor k1. The product generated by this first multiplier 107 is taken to an adder 110, whose output provides the single direction quality value q(x,y)[dx] dependent on the respective direction dx. The second computation unit 104 multiplies the vertical correlation value mv[dx] by the constant weighting factor k3. The multiplication product thus obtained is multiplied by an additional multiplier 109 by the variable vertical weighting factor vg produced by the third computation unit 105, and the resultant product is likewise taken to the adder 110. The weighting factor fg is generated by a decision-making unit 113 as a function of the vertical weighting factor vg, the weighting factor fg being zero, as explained, when the vertical weighting factor is vg=1 or 2. The direction-dependent weighting factor pv[dx] provided at the output of the fourth computation unit 106 is multiplied by a fourth multiplier 111 by the constant weighting factor k4, and the multiplication product thus obtained is multiplied by an additional multiplier 112 with the weighting factor fg. The multiplication product resulting from this is taken to the adder 110.
The adder 110 also receives the additional vertical weighting factor pv[dx], generated by an additional decision-making unit 114 in dependence on the vertical weighting factor vg and the image direction dx being investigated.
Determination of the direction of interpolation presupposes that the composite direction quality values qd(x,y)[dx] found for a pixel being interpolated are compared with each other to find a minimum value for these composite direction quality values. The image direction for which the minimum value is found is then chosen as the direction of interpolation. The number of comparisons required will increase with the number of image directions studied and, thus, with the number of composite direction quality values ascertained.
In order to reduce this expense, one can subsume several image directions into a so-called image direction class and assign one class quality value to each of these direction classes. This procedure is explained below with reference to
Finding a class quality value Q(•) for an image class involves, first, finding for each pixel of the investigated group of pixels a single class quality value Q′(a)(x0,y0). Here, (x+i,y+k) stands for the coordinates of the respective pixel, and C stands for the respective class. The single class quality value Q′(C)(x+i,y+k) is dependent on the single direction quality values q(x+i,y+k)[dx], which was found for the respective pixel (x+i,y+k) for the directions belonging to the image class C. We generally have here:
Q′(C)(x+i,y+k)=f(q(x+i,y+k)[dC] (16)
where f(•) stands for a function of the single direction quality values, dC stands for the different directions of the examined direction class C.
The single class quality value Q′(C)(x+i,y+k) may correspond to the mean value of the single direction quality values found for the particular pixel (x+i,y+k), in which case we have:
where dC stands for the individual image directions assigned to the image class C.
The single class quality value Q′(C)(x+i,y+k) may moreover also correspond to the minimum value of the single direction quality values found for the respective pixel (x+i,y+k), in which case we have:
Q′(C)(x+i,y+k)=min(q(x+i,y+k)[dC]) (16b)
where min(•) stands for the minimum function, and m stands for the number of image directions belonging to the direction class C.
After forming the direction classes, for each of the pixels belonging to the group of pixels there are a number of class quality values Q′(•) available, whose number corresponds to the number of available direction classes. From these single class quality values Q′(•), we then find a class quality value Q(C) for each of the direction classes, which is generally a function of the single class quality values Q′(•). Thus, for this class quality value Q(C) we have:
Q(C)=f(Q′(C)(x+i,y+k)) (17)
where f(•) is any given function of the single class quality values Q′(•).
The class quality value Q(C) for a direction class may correspond, in particular, to the mean value of the single class quality values or to the minimum value of these single class quality values.
Interpolation of an image information value using direction classes may involve the following steps:
In one embodiment, the device 200 also includes grouping arrangements 211, 212, 213, which subsume the single direction quality values into class quality values. The accumulation unit 221 determines in this case from these class quality values a composite class quality value, and the minimum value determining unit 231 determines the minimum value for these class quality values. The minimum value determining unit 231 in this case puts out the value of the class for which the minimum value was found. An image direction determining unit 241 connected to the minimum value determining unit 231 in this case determines within this class the optimal image direction, which is then chosen as the direction of interpolation and communicated to the interpolation unit 251.
Although the present invention has been illustrated and described with respect to several preferred embodiments thereof, various changes, omissions and additions to the form and detail thereof, may be made therein, without departing from the spirit and scope of the invention.
Cited Patent | Filing date | Publication date | Applicant | Title |
---|---|---|---|---|
US5093721 * | Jul 10, 1990 | Mar 3, 1992 | Zenith Electronics Corporation | Line interpolator with preservation of diagonal resolution |
US5625421 | Jan 14, 1994 | Apr 29, 1997 | Yves C. Faroudja | Suppression of sawtooth artifacts in an interlace-to-progressive converted signal |
US6133957 | Oct 14, 1997 | Oct 17, 2000 | Faroudja Laboratories, Inc. | Adaptive diagonal interpolation for image resolution enhancement |
US6181382 * | Jun 2, 1998 | Jan 30, 2001 | Miranda Technologies Inc. | HDTV up converter |
US6731342 * | Jan 5, 2001 | May 4, 2004 | Lg Electronics Inc. | Deinterlacing apparatus and method using edge direction detection and pixel interplation |
US6965705 | Aug 9, 2001 | Nov 15, 2005 | Ndsp, Inc. | Method and system for dynamic angle interpolation in image processing |
US7015971 * | Jun 14, 2002 | Mar 21, 2006 | Stmicroelectronics S.R.L. | Method for converting the scanning format of images, a system and computer program product therefor |
US7023487 * | Jan 25, 2002 | Apr 4, 2006 | Silicon Image, Inc. | Deinterlacing of video sources via image feature edge detection |
US7057664 * | Nov 6, 2002 | Jun 6, 2006 | Broadcom Corporation | Method and system for converting interlaced formatted video to progressive scan video using a color edge detection scheme |
US7092032 * | Sep 22, 2000 | Aug 15, 2006 | Fujitsu General Limited | Scanning conversion circuit |
US7259794 * | Sep 16, 2004 | Aug 21, 2007 | Himax Technologies Limited | De-interlacing device and method therefor |
US7315331 * | Jan 8, 2002 | Jan 1, 2008 | Micronas Gmbh | Method and device for converting video signals |
US7362378 * | Jan 10, 2005 | Apr 22, 2008 | Matsushita Electric Industrial Co., Ltd. | Method of edge based pixel location and interpolation |
US7423691 * | Nov 19, 2001 | Sep 9, 2008 | Matsushita Electric Industrial Co., Ltd. | Method of low latency interlace to progressive video format conversion |
US7474789 * | Mar 22, 2006 | Jan 6, 2009 | Samsung Electronics Co., Ltd. | Method of edge direction detection based on the correlation between pixels of a vector and an edge direction detection system |
US7515205 * | Nov 30, 2006 | Apr 7, 2009 | Pixelworks, Inc. | Weighted absolute difference based deinterlace method and apparatus |
US7538824 * | Aug 18, 2005 | May 26, 2009 | Magnum Semiconductor, Inc. | Systems and methods for reducing noise during video deinterlacing |
US7567294 * | Mar 28, 2005 | Jul 28, 2009 | Intel Corporation | Gradient adaptive video de-interlacing |
US7667773 * | Jun 10, 2005 | Feb 23, 2010 | Samsung Electronics Co., Ltd. | Apparatus and method of motion-compensation adaptive deinterlacing |
US7796191 * | Sep 21, 2005 | Sep 14, 2010 | Nvidia Corporation | Edge-preserving vertical interpolation |
US7808553 * | Oct 12, 2004 | Oct 5, 2010 | Samsung Electronics Co., Ltd. | Apparatus and method for converting interlaced image into progressive image |
US20050134602 | Dec 23, 2003 | Jun 23, 2005 | Lsi Logic Corporation | Method and apparatus for video deinterlacing and format conversion |
US20050157951 | Mar 17, 2005 | Jul 21, 2005 | Sony Corporation | Image processing apparatus and method, recording medium, and program thereof |
US20050225564 | Apr 7, 2005 | Oct 13, 2005 | Pei-Min Shan | Pixel interpolation method and related pixel interpolation system |
US20060268168 * | May 12, 2006 | Nov 30, 2006 | The Hong Kong University Of Science And Technology | Content adaptive de-interlacing algorithm |
US20070070244 * | Mar 22, 2006 | Mar 29, 2007 | Wyman Richard H | Method and system for reducing the appearance of jaggies when deinterlacing moving edges |
EP0550231A1 | Dec 21, 1992 | Jul 7, 1993 | Salon Televisiotehdas Oy | Method for edge-adaptive interpolation of a television picture |
Reference | ||
---|---|---|
1 | De Haan et al., "De-Interlacing of Video Data Using Motion Vectors and Edge Information", Digest of the ICCE '02, Jun. 2002. | |
2 | Doyle et al., "Progressive Scan Conversion Using Edge Information," Proc. of 3rd Int. Workshop on HDTV, Torino, Italy, 1988 pp. 711-721. | |
3 | Fan et al., "Intelligent Intra-Field Interpolation for Motion Compensated Deinterlacing," 3rd International Conference on Information Technology ITRE 2005, Jun. 2005, pp. 200-203. | |
4 | Lee et al., "A New Algorithm for Interlaced to Progressive Scan Conversion Based on Directional Correlations and Its IC Design," IEEE Transactions on Consumer Electronics, vol. 40, No. 2, May 1994, pp. 119-129. | |
5 | T. Doyle, "Interlaced to Sequential Conversion for EDTV Applications," Proc. of 2nd Int. Workshop on Signal Processing of HDTV, L'Aquila, Italy, 1988, pp. 421-430. |
Citing Patent | Filing date | Publication date | Applicant | Title |
---|---|---|---|---|
US8054380 * | Apr 12, 2006 | Nov 8, 2011 | Zoran (France) | Method and apparatus for robust super-resolution video scaling |
US9076230 * | May 9, 2013 | Jul 7, 2015 | Altera Corporation | Circuitry and techniques for image processing |
U.S. Classification | 348/448, 348/452 |
International Classification | H04N11/20, H04N5/44 |
Cooperative Classification | H04N5/14, H04N7/012, G06T3/4007, H04N5/142 |
European Classification | H04N5/14, G06T3/40B, H04N7/01G3 |
Date | Code | Event | Description |
---|---|---|---|
Feb 2, 2007 | AS | Assignment | Owner name: MICRONAS GMBH, GERMANY Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:NIEDERMEIER, ULRICH;REEL/FRAME:018851/0259 Effective date: 20070119 |
May 28, 2010 | AS | Assignment | Owner name: TRIDENT MICROSYSTEMS (FAR EAST) LTD.,CAYMAN ISLAND Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:MICRONAS GMBH;REEL/FRAME:024456/0453 Effective date: 20100408 Owner name: TRIDENT MICROSYSTEMS (FAR EAST) LTD., CAYMAN ISLAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:MICRONAS GMBH;REEL/FRAME:024456/0453 Effective date: 20100408 |
May 3, 2012 | AS | Assignment | Owner name: ENTROPIC COMMUNICATIONS, INC., CALIFORNIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:TRIDENT MICROSYSTEMS, INC.;TRIDENT MICROSYSTEMS (FAR EAST) LTD.;REEL/FRAME:028153/0530 Effective date: 20120411 |
Jan 19, 2015 | FPAY | Fee payment | Year of fee payment: 4 |
May 18, 2015 | AS | Assignment | Owner name: ENTROPIC COMMUNICATIONS, INC., CALIFORNIA Free format text: MERGER AND CHANGE OF NAME;ASSIGNORS:EXCALIBUR ACQUISITION CORPORATION;ENTROPIC COMMUNICATIONS, INC.;ENTROPIC COMMUNICATIONS, INC.;REEL/FRAME:035706/0267 Effective date: 20150430 |
May 19, 2015 | AS | Assignment | Owner name: ENTROPIC COMMUNICATIONS, LLC, CALIFORNIA Free format text: MERGER AND CHANGE OF NAME;ASSIGNORS:ENTROPIC COMMUNICATIONS, INC.;EXCALIBUR SUBSIDIARY, LLC;ENTROPIC COMMUNICATIONS, LLC;REEL/FRAME:035717/0628 Effective date: 20150430 |